Key Advantages
Engineering-grade performance, research-ready from day one.
65 DOF — No Joint Left Behind
Captures every motion dimension: 21 DOF per hand, 9 per arm, 5 waist. Richer data means stronger, more generalizable robot policies.
Millimeter Accuracy at 1000 Hz
65 magnetic absolute encoders with ≤0.5 mm position error and ≤0.1° attitude error — the signal fidelity your algorithms demand.
Full-Stack Dev Integration
C++ / Python / ROS SDK out of the box. Windows & Linux dual support. Compatible with mainstream robotic arm and dexterous hand simulation models.
8–12 h All-Day Battery
Dual 10,000 mAh power banks eliminate charge anxiety. Collect full-day datasets without hardware interruptions.
Ergonomic & Lightweight
Designed for all-day wear at 4.2 kg total. Adjustable gloves (S/M/L) and arm straps fit operators from 160–180 cm.
Configurable Product Lines
Modular DexCap / UL / GL variants let you scale from targeted hand capture to full upper-body teleoperation, adapting to your research scope.
Where DexCap Powers Breakthroughs
From robotics labs to clinical research — wherever human motion data drives discovery, DexCap delivers.
Embodied AI & Robot Imitation Learning
DexCap captures whole-body manipulation data at 65 DOF and 1000 Hz — the fidelity foundation for behavior cloning, Diffusion Policy, and reinforcement learning pipelines targeting humanoid and dexterous-hand platforms.
DexCap captures whole-body manipulation data at 65 DOF and 1000 Hz — the fidelity foundation for behavior cloning, Diffusion Policy, and reinforcement learning pipelines targeting humanoid and dexterous-hand platforms.